| 1 |
>>>> I ran into an out of memory problem. The first mention of it in the |
| 2 |
>>>> kernel log is "mysqld invoked oom-killer". I haven't run into this |
| 3 |
>>>> before. I do have a swap partition but I don't activate it based on |
| 4 |
>>>> something I read previously that I later found out was wrong so I |
| 5 |
>>>> suppose I should activate it. Is fstab the way to do that? I have a |
| 6 |
>>>> commented line in there for swap. |
| 7 |
>> |
| 8 |
>> ... |
| 9 |
>>> |
| 10 |
>>> If you're running any other servers that utilize MySQL like Apache or |
| 11 |
>>> something, check its access logs to see if you had an abnormal number |
| 12 |
>>> of connections. Bruteforce hacking or some kind of flooding/DOS attack |
| 13 |
>>> might cause it to use more memory than it ordinarily would. |
| 14 |
>> |
| 15 |
>> I don't know why I didn't check the apache2 error log before, but I |
| 16 |
>> got the following entry 2 seconds before the server became |
| 17 |
>> unresponsive: |
| 18 |
>> |
| 19 |
>> [error] server reached MaxClients setting, consider raising the |
| 20 |
>> MaxClients setting |
| 21 |
>> |
| 22 |
>> I use the default 256 for MaxClients. This confirms the server was |
| 23 |
>> brought down by too many child processes consuming too much memory. |
| 24 |
>> Looking back at the access_log, it's clear this condition was caused |
| 25 |
>> by the single IP which requested one of my pages about 300 times over |
| 26 |
>> the course of 1 minute. This caused my entire server to lock up for |
| 27 |
>> hours until I rebooted it. |
| 28 |
>> |
| 29 |
>> I hesitate to reduce MaxClients from 256. I think my server should be |
| 30 |
>> able to handle it since it's the default. So I need to prevent my |
| 31 |
>> apache2 child processes from consuming so much memory? apache2 was |
| 32 |
>> restarted about an hour before the lockup so it had a pretty fresh |
| 33 |
>> start. I do use mod_perl which is a memory hog from what I |
| 34 |
>> understand. Do I just need more RAM? |
| 35 |
> |
| 36 |
> Most people do not think about this correctly. Can your server run |
| 37 |
> 10000 Apache processes? No, not enough resources. 1000? No, same problem. |
| 38 |
> 256? I'd say no based on this thread. If you're not going to set it at 10000 |
| 39 |
> why try to keep it at 256? |
| 40 |
|
| 41 |
This is illuminating, thank you. I wonder why the default is set at |
| 42 |
256 though. Do my apache2 processes use an unusual amount of memory |
| 43 |
at 25M-67M? Even at 25M each, you would need 6.4GB of memory for 256 |
| 44 |
processes. Is an apache2 restart the only thing that prevents apache2 |
| 45 |
processes from growing in memory usage indefinitely? |
| 46 |
|
| 47 |
- Grant |
| 48 |
|
| 49 |
|
| 50 |
> Next image a grocery store with 256 checkout lanes, but only four |
| 51 |
> cashiers. Four cashiers trying to run that many lanes is actually slower |
| 52 |
> than having only four lanes. However 32 lanes could faster than 4. People |
| 53 |
> can have their groceries setup, baggers aren't getting in the way, etc. The |
| 54 |
> analogy breaks down a bit, but you get the point. |
| 55 |
> |
| 56 |
> There is no performance gain in configuring for concurrency your |
| 57 |
> hardware and software can not support. |
| 58 |
> |
| 59 |
> kashani |
Archives